American historical myth, especially that of the nineteenth and twentieth centuries, is heavily laden with economic advancement – streets paved with gold and opportunity waiting around every corner. Horatio Alger penned such stories; John D. Rockefeller and Bill Gates embody them. Such anecdotes provide proof for the faithful but hardly a complete chronology of the economic condition of Americans. In reality, how big was the pool of poverty from which the fictional Horatio Alger characters rose? Without a definition of poverty across time, or a count of population therein, the odds of rising above poverty cannot be calculated, and the truth of the myth cannot be tested.

As the United States enters the new millennium, the changing population demographics are dramatic. As of the 2000 U.S.  Census of the Population, African Americans are no longer the single largest minority population – the Hispanic population has pulled even. The percentage of the population that is foreign-born again registers in the double digits. One out of ten Americans lives in a state with a "majority minority” population. How pertinent has the American myth been for demographic minorities in the United States? Which Americans face what odds of being poor?

The implicit social contract of capitalism assigns the individual the responsibility of supporting herself in exchange for the ownership of (or property rights to) the additional value that she produces for society. The incidence of poverty is one measure of how well that contract is succeeding.

Although widely accepted time-series estimates exist for gross national product or consumer prices for periods of over a century, an official measurement of America's poor exists only for the last four decades. This is the first volume of Historical Statistics of the United States to contain statistics on poverty. At the time of the last edition (1975), the U.S. Census Bureau had been publishing national poverty statistics for less than a decade – not long enough to produce an "historical” series. Now, however, the official series of poverty statistics for the United States covers some four decades. Information on poverty in earlier decades, unfortunately, is still limited. The data series in this chapter compile the official poverty statistics covering the period 1959–1999; some earlier unofficial poverty lines and poverty population estimates are also presented.

The data series on poverty in this chapter are presented in two sections: the first documents the dollar levels that define poverty, and the second, who is poor. The first group of poverty tables answers the question, "What are the income levels below which people are classified as poor in the United States?” (Tables Be85-259). The second group answers the question, "Who and how many are poor?” (Tables Be260-411). These tables present poverty counts, poverty rates, and statistics showing the composition of poverty. The poverty count is simply the number of persons or families who are classified as poor. The poverty rate is the prevalence of poverty among select demographic groups (for example, what percentage of whites is poor?). The composition of poverty tells what share of the poor a certain demographic group comprises (for example, what percentage of the poor is white?).

Note that poverty and inequality, while related, are distinct concepts with separate measures and unique historical trends (see the essay in this chapter on inequality). If the poor are defined as those households with incomes below one half the median household income, it is possible for income to be distributed so equally that no one is poor, as historically has been proven by Sweden. Alternatively, defining poverty independent of contemporary living standards can result in almost universal poverty regardless of the income distribution. Such would be the case, for example, if a modern-day poverty line were simply projected backward to the nineteenth century (with appropriate price adjustments). While measures of both poverty and income distribution are certainly considered important socioeconomic barometers of how broadly the benefits of economic production spread, historical income distribution is not addressed in this essay.

Back To Top

Like any definition of poverty, the official U.S. definition of poverty actually has two components – the dollar levels (poverty thresholds) below which people are classified as poor and the definition of income that is compared with those thresholds (Citro and Michael 1995, p. 98). In their present form, the official poverty thresholds for the United States are a matrix of 48 thresholds (dollar figures) that vary by family size, by the number of family members who are children, and (for one- and two-person units only) by the age of the person or family householder.² The Census Bureau uses the matrix of 48 thresholds to tabulate poverty population statistics; the thresholds in the matrix may be referred to as the "detailed thresholds.” However, the figures commonly cited for general purposes are weighted-average poverty thresholds – one for each family size. For three-person families, for instance, the single weighted-average threshold is generally cited in place of the detailed thresholds for a three-adult family, a two-adult/one-child family, and a one-adult/two-child family. The weighting for the average threshold is based on the total number of families of each subtype according to the Current Population Survey (CPS) for the income year in question. Instead of citing all nine weighted-average thresholds, people often cite a single figure to give a general idea of the poverty thresholds; the figure cited is often the weighted-average threshold for a family of four, although more recently, with average family size having dropped, some people cite the weighted-average three-person threshold. The weighted-average poverty thresholds by family size are presented in Table Be95–112. Tables Be177-259 present the full matrices of official poverty thresholds for the "decennial” income years 1959 through 1999.

The U.S. poverty thresholds were originally developed in 1963–1964 by Mollie Orshansky of the Social Security Administration. She began developing the thresholds as part of a Social Security Administration research project; she was not trying to introduce a new general measure of poverty and, indeed, did not know that the Johnson administration was going to initiate a concerted effort against poverty (called the "War on Poverty”) in 1964. In 1965, the Office of Economic Opportunity, the lead agency in the War on Poverty, adopted Orshansky's poverty thresholds as a working or quasi-official definition of poverty. In 1969, the Bureau of the Budget designated the poverty thresholds with some revisions as the federal government's official statistical definition of poverty.

Orshansky based her poverty thresholds on the economy food plan – the cheapest of four food plans developed by the U.S. Department of Agriculture. The Agriculture Department described the economy food plan as being "designed for temporary or emergency use when funds are low.”

Orshansky knew from the Department of Agriculture's 1955 Household Food Consumption Survey, the latest available such survey at the time, that families of three or more persons had spent about one third of their after-tax money income on food in 1955; the one-third figure related to families at all income levels, not just those at lower income levels. Orshansky calculated poverty thresholds for families of three or more persons by taking the dollar costs of the economy food plan for families of those sizes and multiplying the costs by a factor of three – the reciprocal of the one-third food-expense-to-income ratio; this factor of three is known as the "multiplier.” Although there is no standard term for Orshansky's methodology, it can be characterized as a "component-and-multiplier” methodology. In this methodology, she, in effect, scaled down the food/nonfood consumption pattern of a hypothetical average family to the point where food expenditures equaled the cost of the economy food plan and assumed that the family's nonfood expenditures would then be as adequate as the food plan. The methodology did not assume specific dollar amounts for any budget category other than food. Orshansky derived poverty thresholds for two-person families by multiplying the dollar cost of the food plan for that family size by a somewhat higher multiplier (3.7) also derived from the 1955 survey. She calculated the poverty thresholds for one-person units to be directly proportional (at 80 percent) to the thresholds for two-person units.

The base year for the original poverty thresholds was calendar year 1963. In the absence of valid data usable for that purpose, Orshansky did not adjust her thresholds for geographic variations in living costs. She presented her thresholds as a measure of income inadequacy, not of income adequacy – "if it is not possible to state unequivocally ‘how much is enough,’ it should be possible to assert with confidence how much, on average, is too little” (Orshansky 1965, p. 3).

Besides the distinctions retained in the official version of the poverty thresholds (family size, number of members who are children, and aged/nonaged status for smaller units), Orshansky also differentiated her original thresholds by the sex of the family head and by farm/nonfarm status. The family size category ranged from one-person units (unrelated individuals, in Census Bureau terminology) to families of seven or more persons. All of these distinctions resulted in a detailed matrix of 124 poverty thresholds (reduced to 48 in a 1981 revision).

Orshansky developed separate poverty thresholds for farm families because the 1955 survey showed that about 40 percent of the food items consumed by all farm families came from their home farm or garden, rather than being purchased for cash. Another reason for having lower thresholds for farm families involved a technical issue about how farm housing expenses were reported for purposes of determining net farm self-employment income in the CPS. The farm thresholds applied specifically to families living on farms, and not to the broader categories of "rural” or nonmetropolitan families. The farm/nonfarm distinction was eliminated in the 1981 revision of the thresholds.

Orshansky developed separate poverty thresholds for aged one- and two-person units because of policy concerns about the economic status of the aged, most of whom lived in such units. The fact that these thresholds for the aged were lower than those for the nonaged was simply a mechanical consequence of the fact that economy food plan costs for aged persons were lower than those for nonaged adults. Orshansky was not claiming that necessary nonfood expenditures for the aged were or should be lower than those for the nonaged. The aged/nonaged distinction for smaller units has been retained in the official thresholds to the present day.

By late 1965, Social Security Administration policymakers and analysts began to express concern about how to adjust the poverty thresholds for increases in the general standard of living (see the following discussion regarding the income elasticity of the poverty line). In 1968, a Social Security Administration plan to raise the thresholds by 8 percent was rejected, but an interagency Poverty Level Review Committee was initiated to reevaluate the poverty thresholds. It decided to adjust the thresholds only for price changes, and not for changes in the general standard of living. In 1969, the Committee decided that the thresholds would be indexed by the Consumer Price Index (CPI) instead of by the per capita cost of the economy food plan, and that farm poverty thresholds would be set at 85 percent rather than 70 percent of corresponding nonfarm thresholds.⁴ In August 1969, the Bureau of the Budget designated the poverty thresholds with these revisions as the federal government's official statistical definition of poverty. The series of official poverty thresholds given in this chapter therefore represent the same constant-dollar levels as Orshansky's original nonfarm thresholds for the base year 1963. In 1981, in accordance with recommendations of another interagency committee, several minor changes were made in the poverty thresholds, including the elimination of both the farm/nonfarm differential and the distinction between thresholds for female-headed and male-headed families.

Poverty lines are traditionally dichotomized as "absolute” or "relative.” In this context, the phrase "absolute poverty line” is used to denote a poverty definition that is supposedly independent of the living standards, consumption patterns, or development stage of a society.⁵ However, the word "absolute” is commonly used rather loosely in this phrase. It is often unclear whether "absolute” refers to (1) the methodology by which a poverty line is developed, (2) the way in which the poverty line is updated, (3) whether the poverty line has been set at a "subsistence” level based on an implicit ranking of "physical” needs as being more important than "social” needs, or (4) some combination of these. Orshansky's poverty thresholds are often called an absolute poverty line. Because the series of official poverty thresholds represent the same constant-dollar levels as Orshansky's original nonfarm thresholds for the base year 1963, the thresholds are "absolute” only in regard to how they are updated. Note that Orshansky herself described the thresholds she constructed as "relatively absolute” because they were developed from calculations that used the consumption patterns at a particular point in time (1955) of the U.S. population as a whole.

Over the years, a number of ad hoc proposals for changes in one aspect or another of the official U.S. poverty measure have been made; often such proposals have focused quite narrowly on one or another particular feature of the poverty thresholds or the income definition used with them.⁶ In 1995, however, a report was published proposing a new approach for developing a poverty measure that comprised a set of well-thought-out and mutually consistent changes in all major aspects of the poverty measure (Citro and Michael 1995). The report was produced by the Panel on Poverty and Family Assistance appointed in 1992 by the National Research Council in response to a Congressional committee request for a study of the official U.S. poverty measure to provide a basis for its possible revision. The Panel recommended that a new poverty threshold be developed using actual consumer expenditure data on food, clothing, shelter, and utilities (FCSU), with a small amount added for other necessities. The level of the threshold would be within a range based in part on consideration of various expert family budgets and relative and subjective poverty thresholds. The threshold would be updated annually based on changes in median actual consumption expenditures for FCSU. The threshold would be adjusted for different family sizes and types and to reflect geographic variations in housing costs. The Panel would redefine family resources (income) to be consistent with its threshold concept, including money income and the value of near-money benefits (such as food stamps) that can buy goods and services included in the threshold concept, but excluding expenses that cannot be used to buy these goods and services (for example, income and payroll taxes, child care and other work-related expenses, and out-of-pocket medical care costs, including insurance premiums). Illustrative variations of this Panel's recommendations have been published in several Census Bureau reports on experimental poverty measures since 1999 (for example, Short, Garner, et al. 1999; Short 2001). However, no change has been made in the official poverty measure.

Poverty population and poverty rate figures under the Orshansky definition were first published in 1965, but the series was subsequently extended back to 1959, so that there is now some four decades worth of figures (see Figure Be-D). The factor most obviously affecting poverty rates during this period is the business cycle: poverty rates rise during recessions and fall during economic expansions. (During the 1980s and 1990s, poverty rates did not begin to fall until the second or third year of an economic expansion.) Despite the theoretical distinction between poverty and income inequality, the empirical experience with U.S. poverty statistics during the post-1959 period has been that when other factors are held equal, poverty rates increase as income inequality increases (see, for instance, Danziger and Gottschalk 1986).

In 1959, 39.5 million persons – 22.4 percent of all Americans – were in poverty under the official definition in place at that time.⁷ These figures remained essentially unchanged during the next two years, presumably as a result of the 1960–1961 recession. From 1961 to 1969, during the economic expansion of the 1960s, the poverty rate dropped from 21.9 to 12.1 percent; this was the largest percentage point decrease in the official poverty rate over any eight-year span. During the 1969–1970 recession, the poverty rate rose modestly, to 12.6 percent in 1970. It then fell to its all-time low of 11.1 percent – 23.0 million persons – in 1973. During the 1973–1975 recession, the poverty rate rose to 12.3 percent in 1975. As the economy began to expand again, the rate fell to 11.8 percent in 1976, but then remained statistically unchanged through 1979. As a result of two back-to-back recessions (January–July 1980 and July 1981–November 1982), the poverty rate rose from 11.7 percent in 1979 to 15.0 percent in 1982. The poverty rate remained statistically unchanged in 1983 and dropped to 12.8 percent in 1989. During the 1990–1991 recession, the poverty rate rose to 14.2 percent in 1991. As the economy began to expand again, the rate drifted upward to 15.1 percent in 1993, but fell after that, reaching 12.7 percent in 1998.

For the 1959–1998 period as a whole, the overall poverty rate experienced a net decrease of slightly over two fifths – from 22.4 percent in 1959 to 12.7 percent in 1998. All this net decrease had already occurred by 1973, when the poverty rate fell to 11.1 percent.

Poverty rate trends for large demographic groups of the U.S. population are usually broadly similar to those for all persons, being dominated by ups and downs reflecting the business cycle. The major exception to this generalization is the elderly (persons age 65 or older). The elderly experienced a greater net decrease in poverty from 1959 to 1998 than any other group; their poverty rate dropped from 35.2 percent in 1959 to 10.5 percent in 1998, a reduction of more than two thirds. (The corresponding decrease in the number of elderly persons in poverty was from 5.5 million to 3.4 million, with the growth in the overall elderly population partially offsetting the sharp decrease in the elderly poverty rate.) Much of this decrease was related to the Social Security program. In particular, a decrease in the elderly poverty rate from 29.5 percent in 1967 to 15.7 percent (unrevised, versus 14.6 percent revised) in 1974 was closely correlated with Social Security benefit increases during that period (Fisher 1976). After that, several longish periods of relatively little change in the elderly poverty rate were punctuated by shorter periods of decrease: from 15.3 percent in 1981 to 12.4 percent in 1984, and from 12.9 percent in 1992 to 10.5 percent in 1995. As a result of these decreases in the elderly poverty rate, the proportion of the total poverty population who were elderly persons also decreased – from roughly 19 percent in 1966–1970 to just a little less than 10 percent in 1993–1998. The strong effect of the Social Security program on elderly poverty is, of course, the reason why elderly poverty trends are so different from poverty trends in the general population.

The elderly poverty rate was well above both the overall poverty rate and the poverty rate for children (related children younger than age 18 in families) in 1959 and during the 1960s. The elderly poverty rate first fell below the child poverty rate in 1974 and has remained below the latter since then. The elderly poverty rate first fell below the overall poverty rate in 1982 and has remained below the latter since then.

Children (related children younger than age 18 in families) experienced a more modest net decrease in poverty over this period than did the elderly; their poverty rate was 26.9 percent in 1959 and 18.3 percent in 1998, a reduction of just less than one third. The effects of recessions and economic expansions are clearly visible in changes in the child poverty rate. During the 1960s, the poverty rate for children dropped from 26.9 percent in 1959 to 13.8 percent in 1969 – an all-time low that was essentially equaled in 1973. It continued to rise and fall with the business cycle, varying between 15.7 and 22.0 percent during the 1975–1993 period. From 1993 to 1998, the child poverty rate dropped from 22.0 to 18.3 percent. Because the number of children in the general population grew considerably more slowly than the numbers of working-age adults and elderly persons, the proportion of the total poverty population who were children dropped from 43.6 percent in 1959 to 37.3 percent in 1998.

The black population experienced a considerable net decrease in poverty over the 1959–1998 period; the poverty rate for black persons was 55.1 percent in 1959 and 26.1 percent in 1998, a reduction of slightly more than one half. The largest share of this net reduction had occurred by 1973, when the poverty rate for this group fell to 31.4 percent. Despite the large reduction in the poverty rate among African Americans during the 1960s, the proportion of the total poverty population who were black rose from 25.6 percent in 1959 to 32.2 percent in 1973, and then fell back to 26.4 percent in 1998.

The white population also experienced a noticeable net decrease in poverty over the 1959–1998 period; the poverty rate for white persons was 18.1 percent in 1959 and 10.5 percent in 1998, a reduction of slightly more than two fifths. All this net reduction had already occurred by 1973, when the poverty rate for this group fell to 8.4 percent. The proportion of the total poverty population who were white fell from 72.1 percent in 1959 to 65.9 percent in 1973, and then rose slightly to 68.0 percent in 1998.

Poverty statistics for the Hispanic population are not available for years before 1969. Over the 1969–1998 period as a whole, the poverty rate for Hispanic persons showed relatively little net change; it was 24.4 percent in 1969 and 25.6 percent in 1998. The poverty rate for this group showed considerable variation within this period, falling as low as 22 percent in 1973 and 1977–1979 and rising as high as 30–31 percent in 1982 and 1992–1995. Primarily as a result of the growth in the Hispanic population at all income levels, the proportion of the total poverty population who were Hispanic rose from 8.1 percent in 1969 to 23.4 percent in 1998.

Woman-maintained or female-householder families (families with a female householder, no husband present) experienced a modest net decrease in poverty over the 1959–1998 period; their poverty rate was 42.6 percent in 1959 and 29.9 percent in 1998, a reduction of almost one third. Like that of other demographic groups, the female-householder-family poverty rate is affected by economic conditions, although its decreases during economic expansions are more modest than those for some other groups. During the 1960s, the poverty rate for this group dropped from 42.6 percent in 1959 to 32.3 percent in 1968. It rose to 33.9 percent in 1971 and fell to a new low of 30.4 percent in 1979. It rose to 36.3 percent in 1982, fell to 32.2 percent in 1989, and rose to 35.6 percent in 1993. After 1993, it fell fairly steadily, reaching a new all-time low of 29.9 percent in 1998 (although there was no statistically significant difference between this figure and the 1979 figure).

The proportion of all poor families that are female-householder families rose from 23.0 percent in 1959 to 50.3 percent in 1978 and has fluctuated since then between 45.7 and 54.5 percent (see Figure Be-E). This phenomenon has been termed "the feminization of poverty” (Pearce 1978). It is important to note that the feminization of poverty during the 1960s and 1970s took place while the poverty rate for female-householder families was generally falling; this falling poverty rate was more than offset by the growth in the number of female-householder families at all income levels.⁸

Poverty statistics for married-couple families are available only for 1973 and subsequent years. Over the 1973–1998 period, the poverty rate for this group showed no net change; it was 5.3 percent in both 1973 and 1998. (The Census Bureau does have poverty statistics for years before 1973 for the slightly larger group comprising married-couple families plus male-householder families, no wife present – a group formerly termed "male-headed” families; for this group, the poverty rate had dropped from 15.8 percent in 1959 to 5.5 percent in 1973, a reduction of two thirds.)

Poverty rates by employment status back to 1966 were estimated by the Census Bureau through a special tabulation (see Figure Be-F) (Barrington 2000). As expected, for all years the poverty rates were lower for employed persons, and lower still for persons employed full-time and year-round. Interestingly, however, the notable downward trend in poverty since the early 1980s that was present for all persons and all employed persons was not present for persons employed full-time and year-round. This implies that increasing work hours for unemployed or underemployed persons has contributed to reducing poverty among them, whereas the wages of the lowest paid full-time, year-round workers have failed to rise in real terms, thus leaving their poverty rate little changed.

Back To Top

In the century before Mollie Orshansky developed her poverty thresholds, numerous unofficial poverty lines and other measures of income inadequacy were developed by American social workers, labor advocates, government employees, researchers, and others, usually because of concerns about inadequate living standards among working-class and other low-income Americans.¹⁰ Many of these income inadequacy measures were – or were derived from – standard budgets. A standard budget is a list of goods and services that a family of a specified size and composition would need to live at a designated level of well-being, together with the estimated monthly or annual costs of those goods and services.¹¹

Standard budgets and other income inadequacy measures can be developed to represent different standards of living – some of them considerably above poverty. It is thus not always immediately clear which of the pre-Orshansky measures correspond to what Orshansky and others during the 1960s called "poverty” – especially because standardized terminology for classifying living standards was slow to develop. Terms applied to different budgets during the 1900–1920 period included "a fair living wage,” "a fairly proper standard of living,” "a minimum standard,” "safe normal living cost,” and "lowest ‘bare existence.”’ To determine which measures corresponded to Orshansky's poverty concept, we employed a widely used classification scheme developed by Dorothy Douglas and an adaptation of that scheme used by Oscar Ornati (Douglas 1923; Ornati 1966). Historical analysis showed that Douglas's "minimum of subsistence level”¹² and Ornati's "minimum subsistence” level corresponded most closely to Orshansky's poverty concept. Accordingly, we excluded from this chapter more generous standard budgets classified at the level of ‘minimum health and decency,” "minimum adequacy,” or "minimum comfort.”¹³

There is extensive historical evidence from the United States (including but not limited to the minimum subsistence budgets and poverty lines discussed here) and other countries that successive poverty lines develop as "absolute” poverty lines tend to rise in real terms as the real income of the general population increases; this phenomenon is known as the "income elasticity of the poverty line” (Kilpatrick 1973; Fisher 1995; Fisher 1997b, pp. 171–4; see also Citro and Michael 1995, pp. 32, 33, 98–9, 103, 141, and 319). As one way of illustrating this phenomenon, note that when Orshansky presented her poverty thresholds for 1963, she described families living below these thresholds as "lack[ing] the wherewithal to live at anywhere near a tolerable level” (Orshansky 1965, p. 4). But for 1923, an income level (for a family of five) equal in constant dollars to 102 percent of Orshansky's poverty threshold was described as a "comfort level” income "represent[ing] the attainment of the highest class of wage-earners and the cynosure [center of attention or attraction] of the rest.” And for 1907, families with incomes at or above a constant-dollar level equal to 92 percent of Orshansky's five-person threshold were described as "liv[ing] well” and "satisfy[ing]… the reasonable ambitions of an American who puts his life into his work.”¹⁴ Given the extensive American evidence that poverty lines rise in real terms over time as the real income of the general population rises, we generally included in this chapter only pre-Orshansky poverty lines that were developed and used during a particular time period based on the social standards of that time period. We did not include historical "poverty” lines that represent later standards (for example, the official poverty measure) projected back several decades on the basis of price changes only.¹⁵ In two cases, however, we did include measures developed by late twentieth-century scholars for the late nineteenth century based on standards or relevant data from the late nineteenth century.¹⁶

Because of the particularized and nonrecurring nature of most pre-Orshansky poverty lines and minimum subsistence budgets, we have presented them in a special table rather than as a formal data series (see Table Be-G).¹⁷ We have included the source and major defining characteristics of each poverty line. For those poverty lines applicable to a particular family size, we also converted the dollar values to constant dollars and expressed them as percentages of the corresponding Orshansky poverty threshold for 1963. For technical reasons, any such efforts to adjust for price changes over long periods of time are at best rough approximations.

Not all of the pre-Orshansky poverty lines and minimum subsistence budgets in Table Be-G were of equal prominence. Even though the four nineteenth-century figures are of great intrinsic historical interest, none of them were cited by later writers.¹⁸ However, Robert Hunter's 1904 book Poverty had a significant impact on social reformers and others concerned with the problems of poverty; it was even commented on by Britain's H. G. Wells. Hunter's $460/$300 poverty line was the first known poverty line intended to apply to the nation as a whole rather than to a single city or state, and it was mentioned by a number of contemporary and later writers.

From 1906 into the 1920s, a number of standard budgets and other income inadequacy measures were developed. However, much of the public discussion focused on budgets and other figures that were above the minimum subsistence level. Accordingly, some of the most widely cited budgets and other figures during this period – for instance, those in a 1909 study by Robert Chapin – are not included in Table Be-G (Chapin 1909).

Within the limits of the minimum subsistence level, one of the more widely cited measures during the years before the United States entered World War I was the standard budget for a "minimum standard of living” developed by Wood Worcester and Daisy Worthington Worcester, Special Agents (investigators) working for the U.S. Bureau of Labor (later the Bureau of Labor Statistics). Furthermore, the Worcesters' minimum standard/fair standard distinction was adopted by two reviews of budget studies – one published in 1913 by Scott Nearing and the other in 1916 by Maurice Parmelee (see Table Be-H) (Nearing 1913; Parmelee 1916). Similarly, the U.S. Commission on Industrial Relations's "abject poverty” and "decency” standards were adaptations of the Worcesters' two standards (see Table Be-G and Table Be-H).

Several minimum or "bare existence” budgets developed during the 1910s remained in almost total obscurity, but the minimum budget developed by William Ogburn for the National War Labor Board in 1918 was mentioned by several contemporary and later writers; it seems to have been the first budget with which the term "minimum of subsistence level” was used.

Dorothy Douglas's 1923 minimum of subsistence level was by far the most important standard of its type for over a decade. The level and her dollar figures were cited a number of times during the 1920s, while the concept continued to be cited as late as the early 1940s.

Margaret Stecker's emergency budget for the U.S. Works Progress Administration was a minimum subsistence budget cited by various contemporary analysts; in addition, the Textile Workers Union of America issued an upgraded version of this budget in 1944, which they hoped to get accepted as "the lowest conceivable budget… for active, patriotic, self-supporting, self-respecting American workers” in connection with the National War Labor Board's directive to eliminate substandard living conditions among American workers during World War II.

Another New Deal–era figure – the National Resources Committee's $780 figure operationalizing President Roosevelt's "one-third of a nation” quotation – was cited by contemporary and later writers as having been an approximate measure of poverty for this period.

The Congressional Subcommittee on Low-Income Families (SLIF) issued a staff report in 1949 that included a low-income line of $2,000 for families of all sizes for 1948 (see Table Be85–94). During the 1950s, there were not large numbers of Americans writing about poverty and poverty lines. However, among those who did so during the 1949–1958 period (and who did not present poverty lines of their own), the large majority cited the SLIF's $2,000 figure.

In 1958, John Kenneth Galbraith published The Affluent Society. The book included a chapter on poverty. Galbraith's conceptual definition of poverty has often been quoted approvingly by both American and non-American writers.¹⁹ However, his operational definition of poverty ($1,000 for families and unrelated individuals) was criticized as being too low, and was not used by other writers.

In 1958 and 1959, a number of people writing about poverty and poverty lines stopped using the SLIF's $2,000 family low-income line and began using higher figures, such as $3,000. The origin of the $3,000 figure used by some of these writers is not immediately clear.

Unofficial poverty lines were developed and published more frequently during the 1958–1963 period than during the 1949–1958 period. Accordingly, no one poverty line dominated the poverty writings of the early 1960s in the same way that the SLIF's $2,000 figure had dominated the writings of the 1949–1958 period. Robert Lampman published low-income lines for 1947 and 1957 ($2,516 for a family of four in 1957) in a 1959 report done for the Congressional Joint Economic Committee (see Table Be85–94). A number of other writers cited Lampman's low-income line, particularly during 1960. Over the next several years, writers either cited other poverty lines or else cited Lampman's low-income line in conjunction with another – generally a higher – poverty line; the higher poverty lines that were cited included the American Federation of Labor–Congress of Industrial Organizations' $3,000/$1,500 poverty line and the Conference on Economic Progress' $4,000/$2,000 poverty line.

In January 1964, when the War on Poverty was announced, President Johnson's Council of Economic Advisers (CEA) set a poverty line of $3,000 for families and $1,500 for unrelated individuals. This became the United States' first quasi-official poverty line and continued to be so for a little over a year.

In January 1965, Mollie Orshansky published a Social Security Bulletin article presenting her poverty thresholds. In May 1965, as noted earlier, the Office of Economic Opportunity adopted her thresholds as a working definition of poverty; they thus succeeded the CEA's poverty line as the United States' quasi-official poverty line (they were given fully official status as the federal government's statistical definition of poverty in August 1969). The 1965 adoption of Orshansky's thresholds marked the end of the pre-Orshansky period of unofficial poverty lines.

Several estimates of the number of Americans in poverty were done by advocates and analysts before Orshansky; these estimates are presented in Table Be-H because they are of interest to historical scholars and social policy historians. These estimates are too scattered – and were done on too many different bases – to be included among the chapter's data series (if one is looking for a single series of pre-1959 poverty estimates, one will be disappointed.) Table Be-H includes estimates of the total number of Americans in poverty, as well as a few estimates of the proportion of major segments of the population (for example, the urban population, or working families) who were in poverty; it does not include poverty rates for small groups of families in a single city.²⁰ Most of the poverty population estimates in this table were associated with poverty lines or similar measures shown in Table Be-G, although not all of those poverty lines had poverty population estimates associated with them.

As in the case of pre-Orshansky poverty lines, it may be helpful to give a brief assessment of the various pre-Orshansky poverty population estimates and what they can tell us about the incidence of poverty under contemporary definitions.

We found four poverty population estimates for years during the Progressive era and five such estimates for the years 1928–1929. Although each of these two sets of estimates was for years during the same time period, they varied greatly, from about 10 percent to about 50 percent for each set of estimates. The analysts who prepared the individual estimates did the best job that they could to make estimates on the basis of available information, but that information was just too limited. We conclude that their estimates are too weak a basis for us today to make any specific estimate of the poverty rate in the United States (by contemporary standards) during either the Progressive era or the 1928–1929 period. For the period before the mid-1930s, the state of knowledge about the number of Americans in poverty can perhaps best be summed up not in figures but in the words of social insurance theorist and advocate Isaac M. Rubinow:

How much poverty is there in this, the richest country in the world?… Of course, there are no accurate data. We probably know more about the number of poor hogs in this country than the number of poor people. At least, one may have estimates about the hogs. There is a public department to prepare these regularly. There is none to account for the number of poor and dependent. (Rubinow 1929, p. 366)

Only when reasonably good national family income data become available do credible estimates of poverty under contemporary definitions appear. Among the many actions taken by the federal government in response to the Depression were efforts to gather information on the social problems that the New Deal sought to remedy – including low income as well as unemployment. As one government report put it, "Any attempt on the part of Government or business to grapple with basic economic problems must rely heavily on what can be learned of the distribution of income among the various groups of the Nation's consumers” (U.S. National Resources Committee 1938, p. 1). During the mid-1930s, federal agencies conducted two national sample surveys – the Study of Consumer Purchases and the National Health Survey – which collected information on family income as well as other subjects. The addition of limited income questions to the 1940 Decennial Census was presumably motivated by similar concerns. And in March 1940, the U.S. Work Projects Administration (WPA, formerly the Works Progress Administration) initiated the Sample Survey of Unemployment (later the Current Population Survey); when an income supplement was added to this survey in the mid-1940s, it became the principal source of annual national family income data.

One of the earliest survey-based reports with data on the poverty population under a contemporary definition was the preliminary report on the National Health Survey published in 1938 by the U.S. Public Health Service, including data on income in 1935. This report's low-income line was crude, without any adjustment for family size. While it did not aggregate its regional low-income rates into a national rate for all of the 80-plus cities that it covered, it appears that such a national rate would probably have been somewhere in the neighborhood of 45 percent. Including other cities would probably not have changed the national rate greatly, but if the Survey had also included the rural population, the overall national low-income rate for 1935 would probably have been higher – perhaps somewhere between 50 and 60 percent.

A 1942 National Resources Planning Board report included another poverty population estimate based on mid-1930s survey data. The estimate was based on July 1935–June 1936 income data from the Study of Consumer Purchases, combined with some income information from the National Health Survey. This estimate was for urban families only; it thus excluded urban unrelated individuals and rural families and unrelated individuals. It used a poverty line adjusted for family size. It found that 40.0 percent of persons in urban families were either below the WPA emergency budget or in families that had received relief of some kind at some time during 1935–1936. If comparable poverty lines for urban unrelated individuals and the rural population had been available, the overall national poverty rate for 1935–1936 would probably have been higher – perhaps somewhere between 50 and 60 percent.

Intriguingly, we find that extrapolations from these two estimates – using different surveys and different contemporary poverty lines – both suggest that the overall national poverty rate during 1935 or 1935–1936 may perhaps have been in the range of 50 to 60 percent. Even though this estimate is an imprecise range rather than a precise figure, it represents the earliest point at which we can make a statement about the national poverty rate with any confidence.

During the 1990s, Linda Barrington developed an "analytical reconstruction” set of poverty lines for 1939 based on a 1930s Agriculture Department food plan and a multiplier developed from the 1935–1936 Study of Consumer Purchases. She applied these poverty lines to the 1939 earnings (wage and salary) income data collected in the 1940 Decennial Census. As noted in Table Be-H, she came up with an earnings poverty rate of 45.3 percent for persons in 1939. Because the 1940 Census did not collect detailed information on nonearnings money income, this poverty rate is biased upward compared with an unobtainable total-money-income poverty rate for 1939. On the other hand, later analysis showed that Barrington's poverty lines were noticeably lower than contemporary (1930s) minimum subsistence level figures;²¹ in this respect, her poverty rate is biased downward compared with a poverty estimate based on contemporary minimum subsistence figures. All of this suggests that an unobtainable poverty rate for 1939 calculated from total money income using a contemporary minimum subsistence measure might have been somewhere between 45 and 50 percent. On an a priori basis, it seems reasonable to assume that the "actual” poverty rate for 1939 was probably at least a little lower than that for 1935–1936; however, because of differing income definitions in the Study of Consumer Purchases and the 1940 Decennial Census, as well as other data problems, it will probably never be possible to validate that assumption empirically.

Relatively little public attention was paid to the problem of U.S. poverty during World War II. However, in 1949, the SLIF issued a staff report that included a low-income line of $2,000 for families of all sizes for 1948; this $2,000 figure was cited by the majority of those Americans who wrote about poverty and the poverty line during the 1949–1958 period. The SLIF reported that 25 percent of all families were below this line in 1948 and 22 percent of all families were below it in 1954 (after adjustment for inflation). The corresponding poverty rate for persons for 1948 would also have been 25 percent. If poverty rates for persons under this definition had been calculated for other years during the early postwar period, they would have been between 21.6 and 26.3 percent for the years between 1947 and 1952, and between 17.9 and 19.9 percent for the years between 1955 and 1958.

The near-consensus on the $2,000 figure dissolved beginning in 1958. Some writers in that and immediately following years were using poverty lines of $3,000 or even higher (either for all families or for four-person families). Late in 1959, Robert Lampman published a report using $2,516 for a family of four for 1957. For the years between 1957 and 1961, poverty rates were between 17.9 and 20.1 percent under Lampman's definition (as recalculated by Fisher after correcting for minor errors) and in the 20 to 28 percent range under other definitions (under the no-longer-accepted SLIF definition, the poverty rates for those years would have been between 16.5 and 18.3 percent).

In January 1964, when the War on Poverty was announced, President Johnson's Council of Economic Advisers reported that 35 million persons had been below its $3,000/$1,500 poverty line in 1962. The corresponding poverty rate was 19 percent. Under Lampman's definition as recalculated by Fisher, the poverty rate for that year would have been 17.1 percent. Under the official poverty definition used today, the poverty rate for that year – one year before the base year for Orshansky's thresholds – was 21.0 percent.

Estimates and "guesstimates” of the number of Americans in poverty were done as early as 1904. However, it wasn't until after the mid-1930s, when the Great Depression pushed putting people back to work to the top of the national agenda, that income distribution data from surveys and the decennial census became available, making it possible to prepare reliable estimates of the number of persons or families below a specified poverty line. Furthermore, it was not until the 1960s that, taking thresholds originally developed by Mollie Orshansky for her study of children in poverty, the U.S. government adopted an official poverty definition and began measuring poverty systematically. In 1990, during her tenure as U.S. Commissioner of Labor Statistics, Janet Norwood described the poverty rate estimates for the United States as "a data series as sensitive and important as any of the public policy series produced by the federal government” (Norwood 1990, p. 7). Yet, despite this importance, no major overhaul of the poverty measure has been implemented. Embedded still in today's official poverty measurement are American consumption patterns that are almost five decades old. Whether the poverty line in official use today continues in place or is replaced by another poverty line, the way in which poverty is measured will continue both to affect and to be affected by how we view and address the problem of poverty.

Back To Top

Oscar Ornati found a number of standard budgets from the 1905–1960 period (Ornati 1966). Those budgets that he classified as "minimum subsistence” are included in Table Be85–94. Budgets developed for five-person families were recalculated by Ornati for four-person families. Some of these budgets are also listed in Table Be-G. Ornati's budgets are among the best available evidence for what poverty was by contemporary social standards during different years of the pre-1960 period; however, they should not be viewed as a set of national poverty thresholds for that period. They were not promulgated by a single national authority; many of them were for individual cities, not for the nation as a whole; and the figures were never gathered together and presented in one place until Ornati published his book in 1966.

The Subcommittee on Low-Income Families (SLIF) – a specially appointed Congressional subcommittee in 1949–1950 and 1955–1956 – published a low-income line for 1948 in a 1949 staff report and used an unpublished low-income line for 1954 in a 1955 staff report (U.S. Congress 1949, 1955). Robert Lampman published low-income lines for 1947 and 1957 in a 1959 report done for the Joint Economic Committee of Congress, and published low-income lines for 1963 in a 1966 book (Lampman 1959, 1966). Because both the SLIF and Lampman used the CPI to adjust their low-income lines over time, Gordon Fisher was able to calculate figures for both low-income lines for other years during the 1947–1963 period using that index; both original and subsequently calculated figures are shown in Table Be-G (Fisher 1999).

The SLIF's low-income line for 1948 and Lampman's low-income lines for 1957 are generally accepted as analogs for the 1940s and 1950s of Orshansky's poverty thresholds. As can be seen in Table Be-G, Ornati's minimum subsistence budget figures for 1948 and 1957 were (respectively) quite close to the SLIF's 1948 figure and Lampman's 1957 figure; this is the major reason for the conclusion that Ornati's "minimum subsistence” level (rather than one of his higher levels) was the conceptual equivalent of Orshansky's poverty concept.

Figure Be-I shows the constant-dollar equivalents of Ornati's minimum subsistence figures, Lampman's low-income lines, the official poverty thresholds, and three alternative poverty lines. Lampman's low-income lines and the official thresholds are adjusted for price changes only, while the budget figures found by Ornati exhibit the income elasticity of the poverty line – the tendency to rise in real terms as the real income of the general population increases. The constant-dollar decrease in the Ornati budget figures during the 1930s reflects the decrease in the real income of the general population during the Great Depression.

One alternative poverty line presented in Table Be95–112 is the near-poverty threshold for a family of four (125 percent of the weighted-average poverty threshold for a family of four). The Census Bureau began publishing statistics on the number of persons below the near-poverty thresholds in 1970, in response to a recommendation of the interagency Poverty Level Review Committee. These near-poverty thresholds were essentially a successor of a higher set of thresholds that Orshansky developed from the Agriculture Department's low-cost food plan (although they were not calculated in the same way). Some analysts concerned about the level of the official thresholds have cited figures on the number of persons below the near-poverty thresholds. Eligibility for a few federal assistance programs has been set at 125 percent of the poverty guidelines.²²

Another alternative poverty line presented here is the half-of-median-income (relative) poverty line. This definition of poverty was first proposed in the United States by Victor Fuchs.²³ A relative poverty line (unlike a poverty line adjusted only for price changes) changes in real terms as the real standard of living of the society changes, thus reflecting the income elasticity of the poverty line. Although the half-of-the-median income (relative) poverty definition is not used widely in the United States, it has probably been more commonly used than any other non-Orshansky poverty definition.

In almost every year from 1946 through 1992, the American Institute of Public Opinion's Gallup Poll asked the following question: "What is the smallest amount of money a family of four (husband, wife, and two children) needs each week to get along in this community?” The response to this question is known as the "get-along” amount. This amount is one of the major American sources of evidence for the income elasticity of the poverty line, demonstrating that the general population's assessment of a socially acceptable minimum standard of living depends upon the standard of living of the society as a whole. This evidence from the general population complements the evidence from experts in living standards in the form of the standard budgets found by Ornati (Table Be85–94) and shown in Table Be-G.

In contrast to more than 40 years of the get-along question, the Gallup Poll only asked a "poverty” question in one year (1989) using the following wording: "People who have income below a certain level can be considered poor. That level is called the ‘poverty line.’ What amount of weekly income would you use as a poverty line for a family of four (husband, wife and two children) in this community?” Vaughan used the get-along amounts for the period 1947–1989 plus the answer to the 1989 Gallup Poll poverty question to construct a social or "subjective” poverty-line series covering that period.²⁴ That series is presented in Table Be95–112. Vaughan found that his subjective poverty-line series and a relative (half-of-median) poverty-line series were quite close to each other over the whole 1947–1989 period; Vaughan's work and this finding played a significant role in the recommendations for creating a new poverty threshold issued by the National Research Council's Panel on Poverty and Family Assistance (Citro and Michael 1995).

For the "decennial” years 1919, 1939, and 1959–1999, detailed matrices of poverty thresholds are provided in Tables Be113-259.²⁵ The poverty thresholds included in the matrices allow one to determine precisely the official poverty status of an individual or family, given specific family unit characteristics. The matrices for the "decennial” years 1959–1999 present the official poverty thresholds for the United States, adjusted annually for price changes (Tables Be177-259). The matrices for 1919 and 1939 were created by Linda Barrington, duplicating Orshansky's original methodology as closely as possible, but using food plans and multipliers appropriate to the time period of interest (Table Be113–144, Table Be145–176) (Barrington 1997, 1999). Because the Orshansky approach underpins these thresholds, they are comparable in methodology with the official U.S. poverty thresholds. In other words, the 1919 and 1939 matrices are presumably quite close to poverty thresholds that Orshansky would have developed if she had done for the 1910s and the 1930s what she did in the 1960s. Barrington's poverty thresholds take account of householder gender and the farm status of the family, as Orshansky's did, but not whether the householder was 65 years or older. Note that because of the income elasticity of the poverty line, the 1919 and 1939 poverty thresholds so defined are not comparable in level with the official CPI-indexed property thresholds of later years.

Concern over central-city/suburban/nonmetropolitan and regional variation in poverty rates makes the figures in Table Be343–354, Table Be355–370 useful time series. It must be remembered, however, that the underlying official definition of poverty does not adjust for geographic variation in the cost of living. The national poverty thresholds are applied uniformly from El Paso, Texas, to New York City.

The opportunity and the ability of low-income persons to "work their way out of poverty” is an inseparable component of public policy debates over poverty programs and wages. Table Be371–398 provides data on the prevalence of the "working poor” by race. Data are provided for full-time, year-round workers and persons who worked for pay at all during the year. (Note that there is no single official definition of "working poor.”) To be classified as a full-time, year-round worker, one had to be employed at least 35 hours per week and at least 50 weeks a year. While some of the data presented in this table can be found in published U.S. Census Bureau documents, the data necessary to calculate the percentage of full-time, year-round workers in poverty back to 1966 were not available. Custom tabulations were produced by the U.S. Census Bureau from the CPS, March 1967 through March 1999 Supplements, to provide these data (see also Barrington 2000).

Table Be399–411 includes figures on persons below the "near-poverty” threshold (125 percent of the poverty threshold) and those who are "severely poor” (below 50 percent of the poverty threshold). These data are presented by race, Hispanic origin, and age. The aged are highlighted in this table, since comparisons between the aged and the general population give somewhat different results depending on whether one uses the poverty threshold or the near-poverty threshold.

Back To Top

The "unit of analysis” used by the Census Bureau in the process of determining who is in poverty is actually two "units” – families and unrelated individuals. A family is a group of two or more persons related by birth, marriage, or adoption who reside together; all such related persons are considered as members of one family. An unrelated individual is a person 15 years old or older (other than an inmate of an institution) who is not living with any relatives. An unrelated individual may be the only person either living in a housing unit or living in a housing unit in which one or more persons who are not related to the individual in question by birth, marriage, or adoption also live. Poverty thresholds for what have been referred to earlier as one-person units are applied to unrelated individuals. Poverty thresholds for two-person units are applied to two-person families, and so on. The Census Bureau has no short term for the concept "families and/or unrelated individuals.” However, the term "family unit” has been used unofficially to refer to that concept by the poverty guidelines Federal Register notice (since 1978) and by some analysts. "Families and/or unrelated individuals” should not be referred to as "households” because "household” is a separate and distinct concept; as defined by the Census Bureau, a household consists of all the persons who occupy a housing unit, whether they are related to each other or not. A household may thus contain one or more families and/or one or more unrelated individuals; poverty status for each such family and/or unrelated individual is determined separately.

Poverty population figures and poverty rates for persons are presented in most of Tables Be260-411, while poverty population figures and poverty rates for families are limited to Table Be283–309. While correlated, poverty rates for persons and poverty rates for families are distinct statistics and do not always follow the same trend. Trends for the two can vary because of changes in the distribution of families by size, and also because unrelated individuals are included in statistics for all persons but not in statistics for families.

Measuring poverty requires a definition of poverty – a set of poverty thresholds (or lines) and a definition of income that can be compared with those thresholds (see later for the official definition of income). As noted previously, the official poverty thresholds are applied to families and unrelated individuals. The poverty status of a family (or unrelated individual) is determined by comparing the income of the family (or unrelated individual) with the appropriate poverty threshold. If an unrelated individual's income is below the appropriate poverty threshold, he or she is identified as poor. If a family's income (the income of all persons in the family) is below the appropriate poverty threshold, the family and all persons in it are identified as poor. A poverty statistic, such as a poverty count, aggregates the count of poor families or poor persons into a single number. The poverty rate among persons is the percentage of all persons (either unrelated individuals or members of families) within a group who are identified as poor. The poverty rate among families is the percentage of families within a group who are identified as poor. The Census Bureau also prepares poverty statistics for unrelated individuals, but we have not included such statistics here. It should be noted that the poverty rate is a head-count measure – it is calculated simply by dividing the count of the poor by the size of the relevant total population (for example, the total black population in the case of the black poverty rate). The poverty rate thus reflects the prevalence of poverty – how widespread poverty is – but not the severity of poverty.

The severity of poverty is measured by the poverty gap. For a poor family or unrelated individual, the poverty gap is the difference between their actual income and the applicable poverty threshold. The poverty gap is not defined for individual persons within families, although per capita poverty gaps can be calculated for families. Poverty gaps for individual families and unrelated individuals can be aggregated for the whole poverty population and for segments of the total poverty population. The severity of poverty (measured by the aggregate poverty gap) is not reflected in the poverty rate. The Census Bureau dropped the term "poverty gap” from its poverty reports in 1971 and uses the term "income deficit” instead. However, the term "poverty gap” is still understood and used by other analysts.

Poverty statistics for the United States are calculated annually from the CPS. The universe for the CPS includes the civilian noninstitutional population of the United States and members of the armed forces in the United States living off post or with their families on post, but it excludes all other members of the armed forces.²⁶ The "total CPS population” (or "CPS universe”) is derived by subtracting the "civilian institutional population” and "[members of] Armed Forces living without families on post in the United States” from the "total resident (United States) population.”

The category "persons for whom poverty status is determined” – sometimes termed the "poverty universe” – is derived by subtracting "unrelated individuals under age 15” (younger than age 14 before income year 1979) from the "total CPS population.”²⁷ In other words, the poverty universe excludes persons in institutional group quarters, persons in military barracks (members of the armed forces living without families on post in the United States), and unrelated individuals age 15/14.

Because of differences in the questionnaires and data collection procedures (including definition of the poverty universe), estimates of the number of persons below the poverty threshold from the decennial censuses will differ slightly from those derived from the CPS. For example, the number of poor persons in 1979 according to the 1980 Decennial Census was 27.4 million, compared with 26.1 million according to the March 1980 CPS. The comparable figures for families in poverty were 5.7 million and 5.5 million, respectively.²⁸

Under the procedure for measuring poverty by the CPS, persons are surveyed in March of a given year – for instance, in March 1999. At that time, they are asked to report their income from various sources during the preceding calendar year – 1998, in this case. The current year and month's population characteristics are recorded, but the previous year's income is used to determine poverty status for that previous year. Assume, for instance, that a surveyed couple had had a child in February 1999. Their family income in 1998 would be compared with the 1998 poverty threshold for a three-person family (specifically, for a two-adult/one-child family) to determine their 1998 poverty status – even though they were only a two-person family in the year in which the income was received. If it were determined that their 1998 income was below the 1998 poverty threshold for a family of three, their child would be included in the count of poor children in 1998, even though she had not been born yet. In other words, poverty in 1998 is determined on the basis of "income-in-1998-of-persons-as-of-March-1999.” Accordingly, population counts and demographic characteristics associated with poverty statistics for a given year relate to persons as of March of the following year.

The poverty rate indicates what percentage of a demographic group is poor. The composition of poverty indicates what percentage of the poverty population belongs to a certain demographic group. For example, the poverty rate for female-headed families is the percentage of all female-headed families that are poor. The female composition of family poverty is the percentage of all poor families that are female-headed.

The distinction between the poverty rate and the composition of poverty is important. Using a statistical analogy, the poverty rate and the composition of poverty are the opposite conditional probabilities. The impression left by one of these statistics is often quite different from that left by the other. For example, in 1959 the poverty rate among African Americans was 55 percent, but only 25 percent of all poor persons were African Americans. In other words, the majority of African Americans were poor in this year, but the majority of the poor were whites. Time series of both poverty rates and the composition of poverty are included throughout Tables Be260-411.

Definitions of poverty contain two components – the income-based poverty threshold and a definition of income. The definition of income used in the official definition of poverty is before-tax money income. (This definition has been used by the Census Bureau in its annual family income reports since the late 1940s; see the following definition for details.) Accordingly, such statements as "the poverty thresholds do not count noncash benefits” are incorrect; it is the Census Bureau's income definition that does not count noncash benefits.

A family's money income is the sum of the incomes received by all family members ages 15 or older.²⁹ The family's money income is compared with the appropriate poverty threshold to determine the family's poverty status (and thus the poverty status of all persons in the family).

The income definition used to determine poverty status is money income before taxes. In addition to not reflecting personal income taxes that people pay, before-tax money income does not reflect payments for Social Security taxes, union dues, Medicare deductions, and so on, or Earned Income Tax Credit payments that people receive. This income definition also excludes capital gains and noncash benefits from both private and public sources. In other words, money income used to determine poverty status does not reflect the fact that some persons receive such noncash benefits as food stamps, health benefits, rent-free or subsidized housing, goods produced and consumed on the farm, the use of business transportation and facilities, and full or partial payments by business for retirement programs, medical and educational expenses, and so on.

The definition of income that should be used with the poverty thresholds has been a controversial issue for decades. The National Research Council's Panel on Poverty and Family Assistance thoroughly examined this issue in the course of its work. As a result of its examination of the issue, the Panel put great emphasis on the principle that, in poverty measurement, the definition of income used should be consistent with the concept underlying the poverty thresholds. The Panel made a specific recommendation that in developing poverty statistics, any significant change in the definition of income should be accompanied by a consistent adjustment of the poverty thresholds; in particular, the Panel criticized tabulations that add the value of public and private health insurance to families' incomes without adjusting the thresholds to account for medical care needs (Citro and Michael, 1995, pp. 4, 9–10, 37–40, 65–66, 98, 203–206, and 227–231). Because of the Panel's conclusions, we decided not to include figures that made significant changes in the definition of income without making any change in the poverty thresholds.

There have been two revisions in the poverty definition, in 1969 and in 1981. At the time of the 1969 revision, poverty figures for all earlier years back to 1959 were retabulated on the revised basis; accordingly, poverty statistics published since then do not contain any earlier-year figures on the unrevised basis.³⁰ At the time of the 1981 revision, however, the (by now much longer) series for all earlier years was not retabulated; instead, some figures for income year 1980 were retabulated on the revised basis.

Much more frequent than poverty definition revisions have been revisions in the underlying income data series from the CPS. Every decade the Census Bureau introduces statistical controls from the latest decennial census, thus changing the income series. In addition, about once every decade the Census Bureau implements a new computer processing system for the March supplement to the CPS; this also changes the income series. There have been a few other revisions or corrections in the series for other reasons. Information on all these revisions has been gathered from poverty reports in the Current Population Reports series and summarized in the text for Tables Be260-411.

Each time there is a revision in either the income series or poverty definition, the Census Bureau retabulates some figures for the previous income year on the revised basis, so that people can see how the revision affects the poverty population series. Each revision results in greater or lesser changes in poverty population figures and poverty rate figures. In addition, each revision (except the 1981 poverty redefinition) results in changes in total population figures (the numbers of persons and families at all income levels).

Ideally, a presentation of historical poverty statistics would include both unrevised and revised figures for every year for which they are available. In practice, however, we were only able to include one figure for each year for each statistical series. In general, we included the revised figure whenever it was available. In a few cases, however, where both revised and unrevised figures were available for a group but only unrevised figures were available for its components (for example, the child population and its white, black, and Hispanic components), we included the unrevised figure for the whole group as well as for its components.

For any year for which revised figures are available for some statistical series and unrevised figures are available for other statistical series, unrevised figures should not be added to, subtracted from, or divided by revised figures because unrevised figures and revised figures are not strictly comparable with each other.

For some statistical series in these tables, we were able to include poverty figures for 1959 or 1969 from the relevant decennial census when figures were not available for those series for those years from the CPS. For those years, figures from the decennial census should not be added to, subtracted from, or divided by figures from the CPS because the two sets of figures are not strictly comparable with each other.

Back To Top

Series Be272, Series Be276–277, Series Be318, and Series Be335.

Back To Top

Series Be305–306.

Back To Top

Series Be389, Series Be392, and Series Be395.

These series cover persons ages 16 and older.

Back To Top

[TABLE Be-G Poverty lines: 1870–1962]

Back To Top

[TABLE Be-H  Poverty estimates, counts, and rates: 1900–1962]

Back To Top

Series Be85, Series Be91, Series Be102, and Series Be110–112, converted to constant dollars using the consumer price index in series Cc1.

Back To Top

American Federation of Labor–Congress of Industrial Organizations, Department of Research. 1959. "State and Local Tax Burdens Must Be Fairly Shared.” Labor's Economic Review 4 (2): 1ff.

Barrington, Linda. 1997. "Estimating Earnings Poverty in 1939: A Comparison of Orshansky-Method and Price-Indexed Definitions of Poverty.” Review of Economics and Statistics 79 (3): 406–14.

Barrington, Linda. 1999. "Absolutely Relative or Relatively Absolute: The Income Elasticity of Orshansky-Method Poverty Lines, 1919–1959.” Presented at the Annual Meeting of the Economic History Association, Baltimore, October 10, 1999.

Barrington, Linda. 2000. Does a Rising Tide Lift All Boats? America's Full-Time Working Poor Reap Limited Gains in the New Economy, Research Report 1271-00-RR. Conference Board.

Barrington, Linda, and Cecilia A. Conrad. 1994. "At What Cost a Room of Her Own? Factors Contributing to the Feminization of Poverty among Prime-Age Women, 1939–1959.” Journal of Economic History 54 (2): 342–57.

Bernstein, Jared, Chauna Brocht, and Maggie Spade-Aguilar. 2000. How Much Is Enough? Basic Family Budgets for Working Families. Economic Policy Institute.

Brand, Horst. 1960. "Poverty in the United States.” Dissent 7 (4): 334–54.

Bureau of Applied Economics. 1932. Standards of Living: A Compilation of Budgetary Studies, Bulletin number 7, part 2.

Callan, Tim, and Brian Nolan. 1991. "Concepts of Poverty and the Poverty Line.” Journal of Economic Surveys 5 (3): 243–61.

Chapin, Robert Coit. 1909. The Standard of Living among Workingmen's Families in New York City. Charities Publication Committee.

Citro, Constance F., and Robert T. Michael, editors. 1995. Measuring Poverty: A New Approach. National Academy Press.

Comish, Newel Howland. 1923. The Standard of Living: Elements of Consumption. Macmillan.

Conference on Economic Progress. 1962. Poverty and Deprivation in the United States: The Plight of Two-Fifths of a Nation. Conference on Economic Progress.

Council of Economic Advisers. 1954. Economic Report of the President Transmitted to the Congress January 28, 1954. U.S. Government Printing Office.

Council of Economic Advisers. 1955. Economic Report of the President Transmitted to the Congress January 20, 1955. U.S. Government Printing Office.

Council of Economic Advisers. 1964. Economic Report of the President Transmitted to the Congress January 1964 Together With the Annual Report of the Council of Economic Advisers. U.S. Government Printing Office.

Dallas [Texas] Wage Commission, Survey Committee. 1917. Report of Survey Committee to the Dallas Wage Commission and Submitted by Them to the Honorable Mayor and Board of Commissioners of the City of Dallas …, April 25.

Danziger, Sheldon, and Peter Gottschalk. 1986. "Do Rising Tides Lift All Boats? The Impact of Secular and Cyclical Changes on Poverty.” American Economic Review 76 (2): 405–10.

Daugherty, Carroll R. 1938. Labor Problems in American Industry, fourth edition. Houghton Mifflin.

Davis, Allen F. 1963. "The Campaign for the Industrial Relations Commission, 1911–1913.” Mid-America: An Historical Review 45 (4): 211–28.

Douglas, Dorothy W. 1923. "A Description of Standards of Living.” In Paul H. Douglas, Curtice N. Hitchcock, and Willard E. Atkins, editors. The Worker in Modern Economic Society. University of Chicago Press.

Douglas, Paul, and Dorothy Douglas. 1921. "What Can a Man Afford?” American Economic Review, 11 (4), supplement number 2: 1ff.

DuBois, W. E. B. 1967 [1899]. The Philadelphia Negro: A Social Study. Schocken Books.

Early, Frances H. 1983. "The French-Canadian Family Economy and Standard-of-Living in Lowell, Massachusetts, 1870.” In Michael Gordon, editor. The American Family in Social-Historical Perspective, third edition. St. Martin's Press.

Epstein, Lenore A. 1961. "Some Effects of Low Income on Children and Their Families.” Social Security Bulletin 24 (2): 12–17.

Epstein, Lenore A. 1963. "Unmet Need in a Land of Abundance.” Social Security Bulletin 26 (5): 3–11.

Fisher, Gordon M. 1976. "Poverty among the Aged and Social Security Benefits.” In Fred E. Waddell, editor. The Elderly Consumer. Human Ecology Center, Antioch College.

Fisher, Gordon M. 1992. "The Development and History of the Poverty Thresholds.” Social Security Bulletin 55 (4): 3–14.

Fisher, Gordon M. 1995. "Is There Such a Thing as an Absolute Poverty Line over Time? Evidence from the United States, Britain, Canada, and Australia on the Income Elasticity of the Poverty Line.” Manuscript, August.

Fisher, Gordon M. 1997a. "The Development of the Orshansky Poverty Thresholds and Their Subsequent History as the Official U.S. Poverty Measure.” Manuscript, May 1992, revised September 1997.

Fisher, Gordon M. 1997b. "Disseminating the Administrative Version and Explaining the Administrative and Statistical Versions of the Federal Poverty Measure.” Clinical Sociology Review 15: 163–82.

Fisher, Gordon M. 1997c. "From Hunter to Orshansky: An Overview of (Unofficial) Poverty Lines in the United States from 1904 to 1965.” Manuscript, October 1993, revised August 1997.

Fisher, Gordon M. 1998. "Setting American Standards of Poverty: A Look Back.” Focus [newsletter of the Institute for Research on Poverty] 19 (2): 47–52.

Fisher, Gordon M. 1999. "How Many Americans Were Really in Poverty in 1947? Estimates of the U.S. Poverty Population between 1947 and 1963 under Two Contemporary (1949 and 1959) Definitions of Poverty.” Manuscript, October.

Friedman, Rose D. 1965. Poverty: Definition and Perspective. American Enterprise Institute for Public Policy Research.

Fuchs, Victor R. 1965. "Toward a Theory of Poverty.” In Task Force on Economic Growth and Opportunity, editor. The Concept of Poverty. Chamber of Commerce of the United States.

Galbraith, John Kenneth. 1964 [1958]. The Affluent Society. New American Library.

Goldsmith, Selma F. 1962. "Low-Income Families and Measures of Income Inequality.” Review of Social Economy 20 (1): 1–25.

Haines, Michael R. 1981. "Poverty, Economic Stress, and the Family in a Late Nineteenth-Century American City: Whites in Philadelphia, 1880.” In Theodore Hershberg, editor. Philadelphia: Work, Space, Family, and Group Experience in the Nineteenth Century: Essays toward an Interdisciplinary History of the City. Oxford University Press.

Harrington, Michael. 1966 [1962]. The Other America: Poverty in the United States. Penguin Books.

Heller, Walter W. 1964. "Statement… ” (March 17, 1964), pp. 26 ff. in U.S. House of Representatives, Committee on Education and Labor, Economic Opportunity Act of 1964: Hearings before the Subcommittee on the War on Poverty Program… on H.R. 10440, a Bill to Mobilize the Human and Financial Resources of the Nation to Combat Poverty in the United States – Part 1.

Hunter, Robert. 1904. Poverty. Macmillan.

Hunter, Robert. 1908. "Poverty in the United States.” In William D. P. Bliss and Rudolph M. Binder, editors. The New Encyclopedia of Social Reform… . Funk & Wagnalls.

Innes, Judith Eleanor. 1990. Knowledge and Public Policy: The Search for Meaningful Indicators. Transaction Publishers.

Iowa Bureau of Labor Statistics. 1891. Fourth Biennial Report of the Bureau of Labor Statistics for the State of Iowa, 1890–91. G. H. Ragsdale.

Kilpatrick, Robert W. 1973. "The Income Elasticity of the Poverty Line.” Review of Economics and Statistics 55 (3): 327–32.

Kolko, Gabriel. 1962. Wealth and Power in America: An Analysis of Social Class and Income Distribution. Frederick A. Praeger.

Lampman, Robert J. 1959. "The Low Income Population and Economic Growth,” Study Paper number 12. In U.S. Congress, Joint Economic Committee. Study Papers Nos. 12 and 13, The Low Income Population and Economic Growth… The Adequacy of Resources for Economic Growth in the United States… Materials Prepared in Connection with the Study of Employment, Growth, and Price Levels … (December 16, 1959). U.S. Government Printing Office.

Lampman, Robert J. 1966. "Population Change and Poverty Reduction, 1947–75.” In Leo Fishman, editor. Poverty amid Affluence. Yale University Press.

Lauck, W. Jett. 1920. Summary, Analysis and Statement. Presented… on behalf of… the United Mine Workers of America before the United States Anthracite Coal Commission. U.S. Government Printing Office.

Lazarus, Emma. 1989. "The New Colossus.” In The Poems of Emma Lazarus, volume 1. Houghton Mifflin.

Leven, Maurice, Harold G. Moulton, and Clark Warburton. 1934. America's Capacity to Consume. Publication number 56 of the Institute of Economics of the Brookings Institution. Brookings Institution.

[Massachusetts Bureau of Statistics of Labor]. 1871. [Second] Report of the [Massachusetts] Bureau of Statistics of Labor, Embracing the Account of Its Operations and Inquiries from March 1, 1870, to March 1, 1871. Wright & Potter.

More, Louise Bolard. 1907. Wage-Earners' Budgets: A Study of Standards and Cost of Living in New York City. Greenwich House Series of Social Studies – number 1. Henry Holt.

Morgan, James N., Martin H. David, et al. 1962. Income and Welfare in the United States (a study by the Survey Research Center, Institute for Social Research, University of Michigan). McGraw-Hill.

National Industrial Conference Board. 1921. Family Budgets of American Wage-Earners: A Critical Analysis. Research Report number 41. Century Company.

Nearing, Scott. 1913. Financing the Wage-Earner's Family: A Survey of the Facts Bearing on Income and Expenditures in the Families of American Wage-Earners. B. W. Huebsch.

[New York City] Board of Estimate and Apportionment, Bureau of Personal Service. 1917. Report on the Increased Cost of Living for an Unskilled Laborer's Family in New York City. M. B. Brown.

New York State Bureau of Labor Statistics. 1903. Twentieth Annual Report of the [New York State] Bureau of Labor Statistics for the Year Ended September 30, 1902. Argus Company.

Norwood, Janet L. 1990. "Distinguished Lecture on Economics in Government: Data Quality and Public Policy.” Journal of Economic Perspectives 4 (2): 3–12.

Nystrom, Paul H. 1929. Economic Principles of Consumption. Ronald Press.

Ogburn, William F. 1918. "Part I. Digest and Critical Analysis.” In Memorandum on the Minimum Wage and Increased Cost of Living (for members of the National War Labor Board… Submitted by the Secretary [W. Jett Lauck] at the request of the Board at its meeting on July 12, 1918). U.S. Government Printing Office.

Ornati, Oscar, with the editorial assistance of J. Stouder Sweet. 1966. Poverty amid Affluence: A Report on a Research Project Carried Out at the New School for Social Research. Twentieth Century Fund.

Orshansky, Mollie. 1959. "Family Budgets and Fee Schedules of Voluntary Agencies,” Social Security Bulletin 22 (4): 10–17.

Orshansky, Mollie. 1963. "Children of the Poor.” Social Security Bulletin 26 (7): 3–13.

Orshansky, Mollie. 1965. "Counting the Poor: Another Look at the Poverty Profile.” Social Security Bulletin 28 (1): 3–29.

Orshansky, Mollie. 1969. "How Poverty Is Measured.” Monthly Labor Review 92 (2): 37–41.

Parmelee, Maurice. 1916. Poverty and Social Progress. Macmillan.

Pearce, Diana. 1978. "The Feminization of Poverty: Women, Work and Welfare.” Urban and Social Change Review 11 (1 and 2 [combined]) (Special Issue on Women and Work): 28–36.

Plotnick, Robert D., Eugene Smolensky, et al. 2000. "The Twentieth-Century Record of Inequality and Poverty in the United States.” In Stanley Engerman and Robert Gallman, editors. The Cambridge Economic History of the United States, volume 3, The Twentieth Century. Cambridge University Press.

Reed, Louis S. 1933. The Ability to Pay for Medical Care. Publications of the Committee on the Costs of Medical Care, number 25. University of Chicago Press.

Residents of Hull-House, A Social Settlement. … 1895. Hull-House Maps and Papers: A Presentation of Nationalities and Wages in a Congested District of Chicago, Together with Comments and Essays on Problems Growing Out of the Social Conditions. Thomas Y. Crowell.

Reuther, Walter P. 1953. "Report to the Congress of Industrial Organizations.” In 1953 Proceedings of the Fifteenth Constitutional Convention of the Congress of Industrial Organizations, November… 1953. … Congress of Industrial Organizations.

Roosevelt, Franklin D. 1970. "The [Second] Inaugural Address, 1937.” In Harold L. Sheppard, editor. Poverty and Wealth in America. Quadrangle Books.

Ross, Christine, Sheldon Danziger, and Eugene Smolensky. 1987. "The Level and Trend of Poverty in the United States, 1939–1979.” Demography 24 (4): 587–600.

Rubinow, Isaac M. 1929. "Can Private Philanthropy Do It?” Social Service Review 3 (3): 361–94.

Ruggles, Patricia. 1990. Drawing the Line: Alternative Poverty Measures and Their Implications for Public Policy. Urban Institute Press.

Short, Kathleen. 2001. Experimental Poverty Measures: 1999. U.S. Census Bureau, Current Population Reports, series P-60, number 216. U.S. Government Printing Office.

Short, Kathleen, Thesia Garner, et al. 1999. Experimental Poverty Measures: 1990 to 1997. U.S. Census Bureau, Current Population Reports, series P-60, number 205. U.S. Government Printing Office.

Spargo, John. 1968 [1906]. The Bitter Cry of the Children. Quadrangle Books.

Stecker, Margaret Loomis. 1937. Intercity Differences in Costs of Living in March 1935, 59 Cities. Research Monograph 12, Works Progress Administration, Division of Social Research. U.S. Government Printing Office.

Townsend, Peter. 1962. "The Meaning of Poverty.” British Journal of Sociology 13 (3): 210–27.

Tugwell, Rexford Guy, Thomas Munro, and Roy E. Stryker. 1930. American Economic Life and the Means of Its Improvement, third edition. Harcourt, Brace.

U.S. Bureau of the Census. 1971. Current Population Reports, Series P-23, number 37, Social and Economic Characteristics of the Population in Metropolitan and Nonmetropolitan Areas: 1970 and 1960. U.S. Government Printing Office.

U.S. Bureau of the Census. 1973. Census of Population, 1970, Subject Reports, Final Report PC(2)-9A, Low-Income Population. U.S. Government Printing Office.

U.S. Bureau of the Census. 1974. Persons of Spanish Origin in the United States: March 1973. Current Population Reports, series P-20, number 264. U.S. Government Printing Office.

U.S. Bureau of the Census. 1975. Historical Statistics of the United States, Colonial Times to 1970 (Bicentennial Edition), part 1. U.S. Government Printing Office.

U.S. Bureau of the Census. 1980. United States Census of the Population: 1980, volume 1, Characteristics of the Population, part 1, Chapter C. U.S. Government Printing Office.

U.S. Bureau of the Census. [1969 through 1999]. Current Population Reports, series P-60 [selected numbers from number 59 through 207], [selected reports on poverty and money income]. U.S. Government Printing Office.

U.S. Bureau of the Census. Unpublished poverty population tabulations for various years.

[U.S. Bureau of Labor]. 1911. Report on Condition of Woman and Child Wage-Earners in the United States in 19 Volumes,volume 16, Family Budgets of Typical Cotton-Mill Workers, prepared… by Wood F. Worcester and Daisy Worthington Worcester. U.S. Government Printing Office.

U.S. Commission on Industrial Relations. 1916. Industrial Relations: Final Report and Testimony Submitted to Congress…, volume 1. U.S. Government Printing Office.

U.S. Congress, Joint Committee on the Economic Report. 1949. Low-Income Families and Economic Stability – Materials on the Problem of Low-Income Families Assembled by the Staff of the Subcommittee on Low-Income Families… . U.S. Government Printing Office.

U.S. Congress, Joint Committee on the Economic Report. 1955. Characteristics of the Low-Income Population and Related Federal Programs – Selected Materials Assembled by the Staff of the Subcommittee on Low-Income Families… . U.S. Government Printing Office.

U.S. Department of Health, Education, and Welfare. 1976. The Measure of Poverty: A Report to Congress as Mandated by the Education Amendments of 1974. U.S. Government Printing Office.

U.S. House of Representatives. 1916. To Fix the Compensation of Certain Employees: Hearings before the Subcommittee of the Committee on Labor… on H.R. 11876… March… and April… 1916. U.S. Government Printing Office.

U.S. House of Representatives. [1916]. "Minimum Wage Bill… Report.” [To accompany H.R. 11876] 64th Congress, 1st Session, Report number 742.

U.S. National Resources Committee. 1938. Consumer Incomes in the United States: Their Distribution in 1935–36. U.S. Government Printing Office.

U.S. National Resources Planning Board, Committee on Long-Range Work and Relief Policies. 1942. Security, Work, and Relief Policies. U.S. Government Printing Office.

U.S. Public Health Service, National Institute of Health, Division of Public Health Methods. 1938. "The National Health Survey: 1935–1936 – The Relief and Income Status of the Urban Population of the United States, 1935.” Preliminary Reports, National Health Survey, Population Series, Bulletin C.

Vaughan, Denton R. 1993. "Exploring the Use of the Public's Views to Set Income Poverty Thresholds and Adjust Them over Time.” Social Security Bulletin 56 (2): 22–46.

Wyand, Charles S. 1937. The Economics of Consumption. Macmillan.

Back To Top